Recall: Literal objects, arrays
More complex object/array structures
Arrays of objects
Objects with arrays
Extracting data from complex structures
JSON: Complex structures as strings
Saving a data structure to a string
Serialize: JSON.stringify(objOrAry);
Restoring a data structure from a string
Deserialize: JSON.parse(str);
Uses of JSON
"The Fat-Free Alternative to XML" http://www.JSON.org/fatfree.html
"JSON has become the X in Ajax."
web services use JSON to communicate in lieu of XML
(nearly) the same complexity as XML, but with almost zero processing, and immediate data access (no DOM extraction)
client-side code can use JSON to save configuration settings
can be used for arbitrarily large/small amounts of free-form data, on client and server
could be used for storage on server, but most often used to transmit data, not store it
emerging: JSON Schemas (like XML Schemas), data formats in JSON
GeoJSON http://en.wikipedia.org/wiki/GeoJSON
JSON-RPC http://en.wikipedia.org/wiki/JSON-RPC
Advanced uses of JSON
JSONP: "Padded" JSON, fetching a JavaScript file containing JSON from another domain
http://en.wikipedia.org/wiki/JSONP
JsonML: Lossless conversion of XML to JSON
http://jsonml.org/
Some JSON sources
JSON vs XML in Ajax
XML procedure:
1. Fetch XML using Ajax request
2. Browser automatically parses into a DOM object in ajax.responseXML
3. Use getElementsByTagName, .firstChild.nodeValue to isolate desired pieces of information and extract it (long and convoluted process)
JSON procedure:
1. Fetch JSON using Ajax request
2. Run through JSON.parse() to restore object
3. Directly access data in newly restored object
JSON vs XML
<key>value</key>
"key" : "value"
<tag attribute="value">foo</tag>
"key" : {
"attribute": "value",
"content" : "foo"
}
"key" : {
"key" : "value",
"key" : "value"
}
<key>
<key>value</key>
<key>value</key>
</key>
Pros and cons
JSON:
more convenient - direct access to data without having to extract using the DOM
less descriptive/more concise - limited to lists and key/value pairs; some complex XML relations are difficult to represent; tags are more self-describing
no conversion - if the object was created in JavaScript, it's a pain to convert it to XML (or another format) and back again for storage
XML:
more descriptive/less concise - each piece of data is descriptively labeled; elements and attributes can be in more complex structure; some data structures are more intuitively expressed in XML